www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Ch-ch-changes

reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
Hello all,


There's quite a few changes that we're very excited about, that I'd love 
to share to the extent possible.

First, we have decided to extend commit rights to Daniel Murphy and 
Martin Nowak, two heavyweight dmd contributors better known under their 
noms de plume: yebblies and dawgfoto, respectively. Please join me in 
congratulating them for this token of appreciation for their talent and 
hard work.

We want to move dmd forward faster, and we're encouraging committers to 
be more aggressive about reviewing and merging patches. Language changes 
will still have to get through Scylla and Charybdis (that's Walter and 
yours truly), but bug fixes and other non-controversial work can be 
safely parallelized.

Here's a small draft to guide contributors to the compiler proper: 
http://www.prowiki.org/wiki4d/wiki.cgi?Contributing_To_The_D_Compiler. 
It's mostly authored by Don, and we should all add to it sections and 
topics that we consider relevant. At a later point we'll link to the 
document from the website, or integrate it there.

We also want to formalize and automate our processes, including building 
the compiler and its libraries, testing it all, contributing, website, 
and so on. We believe that's a prerequisite to handle (and indeed 
condition) the projected growth of the language. To that end, we'll try 
to define and use a build and release procedure. There's been talk about 
a git workflow; if anyone wants to volunteer creating a detailed 
document describing the steps done, that would be awesome.

Last but not least, we're in talks with a professional conference 
organizer about setting up a D conference. We're looking at some quite 
interesting approaches, but one invariant is that community 
participation and drive is key. We'll get back to you as details firm 
up; for now, lightly hash the months of April and May with a pencil.


Thanks,

Andrei

P.S. Speaking only for myself: there's been robust community growth and 
increase in participation in the past twelve months. It's also clear to 
me that although the resources we have now are fine for today's user 
base, we need to scale well in advance to what we project. By my 
estimates the community size is in the five digits now. To go 1-2 orders 
of magnitude higher, I estimate that continuing to do what we do today 
is far from enough, so we'll need to do some radical changes. Some may 
be risky, and some may be painful. But the as the guy in "Die Hard 2" 
said: no guts, no glory. Let's do this together.
Sep 25 2012
next sibling parent reply "Iain Buclaw" <ibuclaw ubuntu.com> writes:
On Tuesday, 25 September 2012 at 22:09:37 UTC, Andrei 
Alexandrescu wrote:

 Hello all,


 There's quite a few changes that we're very excited about, that 
 I'd love to share to the extent possible.

 First, we have decided to extend commit rights to Daniel Murphy 
 and Martin Nowak, two heavyweight dmd contributors better known 
 under their noms de plume: yebblies and dawgfoto, respectively. 
 Please join me in congratulating them for this token of 
 appreciation for their talent and hard work.

 We want to move dmd forward faster, and we're encouraging 
 committers to be more aggressive about reviewing and merging 
 patches. Language changes will still have to get through Scylla 
 and Charybdis (that's Walter and yours truly), but bug fixes 
 and other non-controversial work can be safely parallelized.

 Here's a small draft to guide contributors to the compiler 
 proper: 
 http://www.prowiki.org/wiki4d/wiki.cgi?Contributing_To_The_D_Compiler. 
 It's mostly authored by Don, and we should all add to it 
 sections and topics that we consider relevant. At a later point 
 we'll link to the document from the website, or integrate it 
 there.


Speaking of the 'front-end proper':
"Given that all back-ends share the same front-end, you should 
expect the same compatibility level across all compilers for a 
given compiler release."

Unfortunately, this is not quite the case yet. I'm in the middle 
of drafting something, but the ultimate goal is to get us all 
working together (dmd, gdc, ldc) so that there can genuinely be a 
shared, portable source base for the D front-end, used by all 
maintainers, without the neccesity to modify the original code, 
or use conditionals based on which compiler it's used in. eg: 
#ifdef IN_GCC, IN_LDC, IN_DMD...


Keep up the good work! :~)

Regards,
Iain.
Sep 25 2012
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 9/25/12 6:31 PM, Iain Buclaw wrote:

 Speaking of the 'front-end proper':
 "Given that all back-ends share the same front-end, you should expect
 the same compatibility level across all compilers for a given compiler
 release."

 Unfortunately, this is not quite the case yet. I'm in the middle of
 drafting something, but the ultimate goal is to get us all working
 together (dmd, gdc, ldc) so that there can genuinely be a shared,
 portable source base for the D front-end, used by all maintainers,
 without the neccesity to modify the original code, or use conditionals
 based on which compiler it's used in. eg: #ifdef IN_GCC, IN_LDC, IN_DMD...


That's great, keep us posted.


 Keep up the good work! :~)


Well right back atcha.


Andrei
Sep 25 2012
prev sibling next sibling parent Ben Davis <entheh cantab.net> writes:
Seems like as good a time as any to say how awesome I think D is, and 
how much the world needs it! Where I work, we have a monstrosity of a 
project using all of Java, C++ and Lua, with all three interoperating in 
all directions. We've been asked to write all new code in C++ for 
reasons of portability. Having used Java extensively and D a little, I. 
Cannot. Stand. The. Pain. Of. Need for header files, lack of automatic 
memory initialisation, lack of decent IDE support because the language 
is just that bad... So while they're right about Java not being 
portable, I'm holding out and continuing to write in Java anyway. :)

So I absolutely want D to grow and get the support it needs to be able 
to replace C++, and the sooner the better! Go D! :)

(Normally I'd worry how many people I just offended, but probably not 
here :D)

On 25/09/2012 23:10, Andrei Alexandrescu wrote:

 Hello all,


 There's quite a few changes that we're very excited about, that I'd love
 to share to the extent possible.

 First, we have decided to extend commit rights to Daniel Murphy and
 Martin Nowak, two heavyweight dmd contributors better known under their
 noms de plume: yebblies and dawgfoto, respectively. Please join me in
 congratulating them for this token of appreciation for their talent and
 hard work.

 We want to move dmd forward faster, and we're encouraging committers to
 be more aggressive about reviewing and merging patches. Language changes
 will still have to get through Scylla and Charybdis (that's Walter and
 yours truly), but bug fixes and other non-controversial work can be
 safely parallelized.

 Here's a small draft to guide contributors to the compiler proper:
 http://www.prowiki.org/wiki4d/wiki.cgi?Contributing_To_The_D_Compiler.
 It's mostly authored by Don, and we should all add to it sections and
 topics that we consider relevant. At a later point we'll link to the
 document from the website, or integrate it there.

 We also want to formalize and automate our processes, including building
 the compiler and its libraries, testing it all, contributing, website,
 and so on. We believe that's a prerequisite to handle (and indeed
 condition) the projected growth of the language. To that end, we'll try
 to define and use a build and release procedure. There's been talk about
 a git workflow; if anyone wants to volunteer creating a detailed
 document describing the steps done, that would be awesome.

 Last but not least, we're in talks with a professional conference
 organizer about setting up a D conference. We're looking at some quite
 interesting approaches, but one invariant is that community
 participation and drive is key. We'll get back to you as details firm
 up; for now, lightly hash the months of April and May with a pencil.


 Thanks,

 Andrei

 P.S. Speaking only for myself: there's been robust community growth and
 increase in participation in the past twelve months. It's also clear to
 me that although the resources we have now are fine for today's user
 base, we need to scale well in advance to what we project. By my
 estimates the community size is in the five digits now. To go 1-2 orders
 of magnitude higher, I estimate that continuing to do what we do today
 is far from enough, so we'll need to do some radical changes. Some may
 be risky, and some may be painful. But the as the guy in "Die Hard 2"
 said: no guts, no glory. Let's do this together.
Sep 25 2012
prev sibling next sibling parent "bearophile" <bearophileHUGS lycos.com> writes:
Andrei Alexandrescu:


 I estimate that continuing to do what we do today is far from 
 enough, so we'll need to do some radical changes. Some may be 
 risky, and some may be painful. But the as the guy in "Die Hard 
 2" said: no guts, no glory. Let's do this together.


Currently the D community is still small, so I suggest to add 
this link to the D home page:
http://rosettacode.org/wiki/Category:D

Bye,
bearophile
Sep 25 2012
prev sibling next sibling parent reply Jacob Carlborg <doob me.com> writes:
On 2012-09-26 00:10, Andrei Alexandrescu wrote:

 Hello all,


 There's quite a few changes that we're very excited about, that I'd love
 to share to the extent possible.

 First, we have decided to extend commit rights to Daniel Murphy and
 Martin Nowak, two heavyweight dmd contributors better known under their
 noms de plume: yebblies and dawgfoto, respectively. Please join me in
 congratulating them for this token of appreciation for their talent and
 hard work.


That's awesome.


 We want to move dmd forward faster, and we're encouraging committers to
 be more aggressive about reviewing and merging patches. Language changes
 will still have to get through Scylla and Charybdis (that's Walter and
 yours truly), but bug fixes and other non-controversial work can be
 safely parallelized.


Sounds good.


 Here's a small draft to guide contributors to the compiler proper:
 http://www.prowiki.org/wiki4d/wiki.cgi?Contributing_To_The_D_Compiler.
 It's mostly authored by Don, and we should all add to it sections and
 topics that we consider relevant. At a later point we'll link to the
 document from the website, or integrate it there.


This should really be moved to a "Contribute" section on dlang.org. I've 
been thinking about writing that section a couple of times but I have 
not done so yet.

-- 
/Jacob Carlborg
Sep 25 2012
parent reply "Nick B" <nick.barbalich gmail.com> writes:
On Wednesday, 26 September 2012 at 06:28:00 UTC, Jacob Carlborg 
wrote:

 On 2012-09-26 00:10, Andrei Alexandrescu wrote:



 By my estimates the community size is in the five digits now.


So how do you estimate these numbers ?

Nick B
Sep 26 2012
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 9/26/12 5:37 AM, Nick B wrote:

 On Wednesday, 26 September 2012 at 06:28:00 UTC, Jacob Carlborg wrote:

 On 2012-09-26 00:10, Andrei Alexandrescu wrote:



 By my estimates the community size is in the five digits now.


 So how do you estimate these numbers ?

 Nick B


Downloads, website visits, newsgroup visits. I want us to get a lot more 
transparent about that all, soon.

Andrei
Sep 26 2012
prev sibling next sibling parent reply "monarch_dodra" <monarchdodra gmail.com> writes:
On Tuesday, 25 September 2012 at 22:09:37 UTC, Andrei 
Alexandrescu wrote:

 Hello all,

 [SNIP]

 We want to move dmd forward faster, and we're encouraging 
 committers to be more aggressive about reviewing and merging 
 patches. Language changes will still have to get through Scylla 
 and Charybdis (that's Walter and yours truly), but bug fixes 
 and other non-controversial work can be safely parallelized.

 [SNIP]

 Thanks,

 Andrei

 [SNIP]


Any chance we could see some sort of "dedicated hand holders" for 
the newer contributors? That or "semi-trusted" validators: They 
wouldn't have the power to merge themselves, but could still tag 
pulls as "Validated by XXX": Basically, a screening process.

I know that personally, I feel like I can help, but at the same 
time, I know I sometimes try to push some bad code/design. I'm 
glad there is someone that actually takes the time to review my 
code, but at the same time, I don't want to bother the dedicated 
pullers that already have a lot on their plate.

Also (but more related to Phobos), I have myself tried to review 
others' code (when my field of knowledge permits), but it seems 
like it is not something that is done very often.

Overall, I think the most important thing would be promoting 
cross puller reviews. If only just to say "looks good to me!": It 
can give the puller that extra confidence in the pull, without 
having to delve into it.

I don't mind having to wait 3 weeks for a pull to go through, but 
it is kind of weird when the pull just sits there with no 
comments *from anyone* during those 3 weeks...
Sep 26 2012
parent reply =?UTF-8?B?QWxleCBSw7hubmUgUGV0ZXJzZW4=?= <alex lycus.org> writes:
On 26-09-2012 13:47, monarch_dodra wrote:

 On Tuesday, 25 September 2012 at 22:09:37 UTC, Andrei Alexandrescu wrote:

 Hello all,

 [SNIP]

 We want to move dmd forward faster, and we're encouraging committers
 to be more aggressive about reviewing and merging patches. Language
 changes will still have to get through Scylla and Charybdis (that's
 Walter and yours truly), but bug fixes and other non-controversial
 work can be safely parallelized.

 [SNIP]

 Thanks,

 Andrei

 [SNIP]


 Any chance we could see some sort of "dedicated hand holders" for the
 newer contributors? That or "semi-trusted" validators: They wouldn't
 have the power to merge themselves, but could still tag pulls as
 "Validated by XXX": Basically, a screening process.


There are several people who go over pull requests and weigh in on them 
even though they are not committers (me being one). Anyone can feel free 
to do so. :)


 I know that personally, I feel like I can help, but at the same time, I
 know I sometimes try to push some bad code/design. I'm glad there is
 someone that actually takes the time to review my code, but at the same
 time, I don't want to bother the dedicated pullers that already have a
 lot on their plate.


 From what I've seen, it's not like your pull requests have been 
completely unreasonable or annoying to review or anything.

I see it as the responsibility of committers/authors to review pull 
requests that touch areas they're familiar with, so long as their spare 
time permits it.


 Also (but more related to Phobos), I have myself tried to review others'
 code (when my field of knowledge permits), but it seems like it is not
 something that is done very often.


Yes, we desperately need more people to do reviews.


 Overall, I think the most important thing would be promoting cross
 puller reviews. If only just to say "looks good to me!": It can give the
 puller that extra confidence in the pull, without having to delve into it.

 I don't mind having to wait 3 weeks for a pull to go through, but it is
 kind of weird when the pull just sits there with no comments *from
 anyone* during those 3 weeks...


It's quite unfortunate that pull request reviews take so long in our 
community, and it really is something we need to fix. But without people 
actually willing to do reviews, we're not going to get anywhere.

Perhaps a post in D.announce that encourages peer review would be in order.

-- 
Alex Rønne Petersen
alex lycus.org
http://lycus.org
Sep 26 2012
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 9/26/12 11:19 AM, Alex Rønne Petersen wrote:

 On 26-09-2012 13:47, monarch_dodra wrote:

 Any chance we could see some sort of "dedicated hand holders" for the
 newer contributors? That or "semi-trusted" validators: They wouldn't
 have the power to merge themselves, but could still tag pulls as
 "Validated by XXX": Basically, a screening process.


 There are several people who go over pull requests and weigh in on them
 even though they are not committers (me being one). Anyone can feel free
 to do so. :)


I should add that I've always found your comments and LGTMs very useful. 
I meant to write that in reply to one of them, but forgot. Thanks!



Andrei
Sep 26 2012
prev sibling next sibling parent Dmitry Olshansky <dmitry.olsh gmail.com> writes:
On 26-Sep-12 02:10, Andrei Alexandrescu wrote:

 First, we have decided to extend commit rights to Daniel Murphy and
 Martin Nowak, two heavyweight dmd contributors better known under their
 noms de plume: yebblies and dawgfoto, respectively. Please join me in
 congratulating them for this token of appreciation for their talent and
 hard work.


Congrats!


 We want to move dmd forward faster, and we're encouraging committers to
 be more aggressive about reviewing and merging patches. Language changes
 will still have to get through Scylla and Charybdis (that's Walter and
 yours truly), but bug fixes and other non-controversial work can be
 safely parallelized.


*Me appreciatively nods*



 We also want to formalize and automate our processes, including building
 the compiler and its libraries, testing it all, contributing, website,
 and so on. We believe that's a prerequisite to handle (and indeed
 condition) the projected growth of the language. To that end, we'll try
 to define and use a build and release procedure. There's been talk about
 a git workflow; if anyone wants to volunteer creating a detailed
 document describing the steps done, that would be awesome.


Risking a shameless plug, a prototype of the D changelog generator 
written in D:
https://gist.github.com/3734045
(pulls resolved issues from bugzilla so that nothing is left behind)

I've posted it before but it seems to have been lost in Phobos mailing 
list. The Q is: if we need such a tool and more generally what kind of 
tools/scripts could help the release process to go smoother.

Ideally I think it should get close to the point of "a single hit of a 
button" to run all relevant tasks, checks and uploads.

In other words - core developers what's on your wishlist?


 Last but not least, we're in talks with a professional conference
 organizer about setting up a D conference. We're looking at some quite
 interesting approaches, but one invariant is that community
 participation and drive is key. We'll get back to you as details firm
 up; for now, lightly hash the months of April and May with a pencil.


Cool.


-- 
Dmitry Olshansky
Sep 26 2012
prev sibling next sibling parent reply "nazriel" <spam dzfl.pl> writes:
On Tuesday, 25 September 2012 at 22:09:37 UTC, Andrei 
Alexandrescu wrote:

 Hello all,


 There's quite a few changes that we're very excited about, that 
 I'd love to share to the extent possible.

 First, we have decided to extend commit rights to Daniel Murphy 
 and Martin Nowak, two heavyweight dmd contributors better known 
 under their noms de plume: yebblies and dawgfoto, respectively. 
 Please join me in congratulating them for this token of 
 appreciation for their talent and hard work.


This is great news!

Congratulations Daniel and Martin!
Sep 26 2012
next sibling parent reply Philippe Sigaud <philippe.sigaud gmail.com> writes:
On Tuesday, 25 September 2012 at 22:09:37 UTC, Andrei Alexandrescu wrote:

 Hello all,


 There's quite a few changes that we're very excited about, that I'd love
 to share to the extent possible.

 First, we have decided to extend commit rights to Daniel Murphy and Martin
 Nowak, two heavyweight dmd contributors better known under their noms de
 plume: yebblies and dawgfoto, respectively. Please join me in congratulating
 them for this token of appreciation for their talent and hard work.


You guys are great! Keep up the good work!
Sep 26 2012
parent reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 9/26/12 6:16 PM, Philippe Sigaud wrote:

 On Tuesday, 25 September 2012 at 22:09:37 UTC, Andrei Alexandrescu wrote:

 Hello all,


 There's quite a few changes that we're very excited about, that I'd love
 to share to the extent possible.

 First, we have decided to extend commit rights to Daniel Murphy and Martin
 Nowak, two heavyweight dmd contributors better known under their noms de
 plume: yebblies and dawgfoto, respectively. Please join me in congratulating
 them for this token of appreciation for their talent and hard work.


 You guys are great! Keep up the good work!


You are great, too, and would be even greater if you finalized your 
parser generator and submitted it to Phobos. Most of the questions I get 
asked at conferences are about your parser and Dmitry's regex.

Andrei
Sep 26 2012
parent reply Philippe Sigaud <philippe.sigaud gmail.com> writes:
On Thu, Sep 27, 2012 at 1:30 AM, Andrei Alexandrescu
<SeeWebsiteForEmail erdani.org> wrote:


 You are great, too, and would be even greater if you finalized your parser
 generator and submitted it to Phobos. Most of the questions I get asked at
 conferences are about your parser and Dmitry's regex.


Thanks. That's why I keep having people forking/starring it on github
:) And that's why I'm working on it regularly: I recently upgraded its
speed by one to two orders of magnitude (almost two in fact) at
compile- and runtime, got it linear by using memoization, added
introspection capabilities and grammar diagnostic. I also corrected
numerous bugs and maybe found a way to deal with left-recursive rules
which the D grammar is so fond of.
I'm now using it to get a pattern-matching engine in D: it's heavily
related to the recent tuple thread, where many people ask for
deconstruction, only I'm doing it for any type and range.

Philippe
Sep 26 2012
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 9/27/12 1:00 AM, Philippe Sigaud wrote:

 On Thu, Sep 27, 2012 at 1:30 AM, Andrei Alexandrescu
 <SeeWebsiteForEmail erdani.org>  wrote:


 You are great, too, and would be even greater if you finalized your parser
 generator and submitted it to Phobos. Most of the questions I get asked at
 conferences are about your parser and Dmitry's regex.


 Thanks. That's why I keep having people forking/starring it on github
 :) And that's why I'm working on it regularly: I recently upgraded its
 speed by one to two orders of magnitude (almost two in fact) at
 compile- and runtime, got it linear by using memoization, added
 introspection capabilities and grammar diagnostic. I also corrected
 numerous bugs and maybe found a way to deal with left-recursive rules
 which the D grammar is so fond of.
 I'm now using it to get a pattern-matching engine in D: it's heavily
 related to the recent tuple thread, where many people ask for
 deconstruction, only I'm doing it for any type and range.

 Philippe


This is a 50 shades beyond awesome, and has "Phobos submission before 
the end of 2012 (and the world)" written all over it!

Andrei
Sep 27 2012
prev sibling parent "Jonathan M Davis" <jmdavisProg gmx.com> writes:
On Thursday, September 27, 2012 00:16:41 Philippe Sigaud wrote:

 On Tuesday, 25 September 2012 at 22:09:37 UTC, Andrei Alexandrescu wrote:

 Hello all,
 
 
 There's quite a few changes that we're very excited about, that I'd love
 to share to the extent possible.
 
 First, we have decided to extend commit rights to Daniel Murphy and Martin
 Nowak, two heavyweight dmd contributors better known under their noms de
 plume: yebblies and dawgfoto, respectively. Please join me in
 congratulating them for this token of appreciation for their talent and
 hard work.


 You guys are great! Keep up the good work!


Indeed. We're lucky to have them.

- Jonathan M Davis
Sep 26 2012
prev sibling parent deadalnix <deadalnix gmail.com> writes:
Le 26/09/2012 00:10, Andrei Alexandrescu a écrit :

 There's quite a few changes that we're very excited about, that I'd love
 to share to the extent possible.

 First, we have decided to extend commit rights to Daniel Murphy and
 Martin Nowak, two heavyweight dmd contributors better known under their
 noms de plume: yebblies and dawgfoto, respectively. Please join me in
 congratulating them for this token of appreciation for their talent and
 hard work.


Congratulations ! You deserve it.
Sep 27 2012